Understanding Large Language Models: A Beginner's Guide
A ground-up explanation of how LLMs work — architecture, training, and the intuition behind why they generate coherent text.
Feb 2026 · Large Language Models · Generative AI
Blog
AI & Machine Learning Writing — LLMs, RAG, PyTorch, and Production ML
Technical articles on building and understanding AI systems — from transformer internals and generative AI to production RAG pipelines and machine learning fundamentals. Published on Medium.
How GPT-2 Works: Building a Transformer Language Model From Scratch in PyTorch
Step-by-step implementation of GPT-2 in PyTorch — attention heads, positional encoding, and the full autoregressive training loop.
Apr 2025 · PyTorch · Transformers · Deep Learning
Building an ML-Powered Survey Analysis System with RAG: A Deep Dive
How to combine retrieval-augmented generation with ML classifiers to analyze open-ended survey responses at scale.
Oct 2024 · RAG · NLP · Production ML
OpenFPL-Scout-AI: Data-Driven Fantasy Premier League Success
How ML models, time-series features, and live data APIs power automated FPL team selection and player performance prediction.
Aug 2025 · Machine Learning · Predictive Analytics
Master Fantasy Premier League with FantasyLTX
A data-driven approach to FPL success using machine learning models for player selection, captaincy decisions, and transfer strategy.
Sep 2024 · Machine Learning · Data Analysis
Image Similarity Estimation Using a Siamese Network with Triplet Loss
Practical guide to building a Siamese network in PyTorch with triplet loss for learning image embeddings and measuring visual similarity.
Aug 2024 · PyTorch · Computer Vision · Deep Learning
Building a K-Nearest Neighbors Classifier from Scratch
Comprehensive guide to implementing KNN in Python from first principles — distance metrics, decision boundaries, and performance trade-offs.
Aug 2024 · Machine Learning · Python · Algorithms
Building a Naive Bayes Classifier from Scratch
Step-by-step construction of a Naive Bayes classifier using Bayes' theorem and conditional probability — no ML library required.
Aug 2024 · Machine Learning · Python · Classification
Building a Simple Logistic Regression Model from Scratch with NumPy
Binary classification from the ground up — gradient descent, sigmoid activation, and loss optimization using only NumPy.
Jul 2024 · Machine Learning · NumPy · Deep Learning
Understanding Linear Regression by Building It from Scratch with NumPy
Foundational ML — implementing linear regression using the normal equation and gradient descent, with visual explanations of the math.
Jul 2024 · Machine Learning · NumPy · Fundamentals